Does AI Drive Bookings? Real Numbers from Three of Our Hotels
782 AI sessions, 30 days, three Hungarian hotels: we show how often today's AI-referred guest books compared with organic traffic, and why last-click underestimates the impact. Real GA4 data, with methodology.
Most writing about AI visibility is about promises. Here come the numbers. We lined up the real Google Analytics data from the past 30 days across three of our hotels: how many visitors came from AI assistants (ChatGPT, Perplexity, Gemini, Copilot), how many from organic search, and which channel drove bookings at what rate.
First, the most important point: AI traffic is still small today. Anyone who claims otherwise either isn't measuring or is trying to sell you something. But "small" doesn't mean "doesn't matter". It means now is the moment to claim position cheaply where the curve is starting to bend upward.
The gist in brief
- Across our three hotels, 782 sessions arrived from AI assistants over 30 days. Alongside the 11,785 organic sessions, that's ~6% of search traffic.
- The booking conversion of AI-referred visitors ranged between 0.4–1.3%, versus 1.5–2% for organic: today's AI guest is still researching more than booking.
- Last-click measurement underestimates AI's impact: someone who first noticed you in ChatGPT often comes back the next day — searching for your name or arriving directly — to book.
The numbers — per hotel and combined
| AI sessions (30 days) | AI booking rate | Organic booking rate | |
|---|---|---|---|
| Hotel "A" | 468 | 0.43% | 1.47% |
| Hotel "B" | 238 | 1.26% | 2.02% |
| Hotel "C" | 76 | 1.32% | 1.82% |
| Combined | 782 | 0.77% | 1.78% |
The booking rate here refers strictly to the web analytics booking event, per session — not an enquiry, not a newsletter sign-up, but an actual booking.
Two things are immediately clear. One: AI traffic exists and is measurable. It's not a forecast but several hundred visitors a week, already today, at Hungarian hotels. Two: the AI-referred visitor still books on the first visit less often than organic. That makes sense too: the guest arriving from ChatGPT is often still in the research phase — comparing, looking around, then deciding later.
Why last-click underestimates it — and why that matters
The figures above are so-called last-click data: they attribute the booking to the session in which it happened. But AI's effect typically doesn't materialise in that session. The typical path: the guest asks ChatGPT in the evening, gets three hotel names, the next day searches for your name on Google (or types your address directly), and books then. In the measurement, that booking gets an "organic/brand" or "direct" label — even though AI made the recommendation.
AI primarily drives decisions, not clicks. The click is often booked elsewhere.
That's why 0.77% is a lower bound for the real AI impact. For the full picture you also need to watch the trend in brand searches: as your AI visibility grows, searches for your name typically grow too.
A surprise in the margins: Bing
While everyone watches ChatGPT, our numbers hold a quiet surprise: Bing organic traffic is small but converts exceptionally well — 2.3–2.8% depending on the hotel, and even higher at one of them. The reason is obvious: Bing is partly backed by the same Microsoft ecosystem as Copilot, and Bing users tend to be older, more affluent and closer to a decision. The lesson: the "AI era" isn't a single channel. It's worth measuring the whole map.
How to measure the same thing at your own hotel
You don't need separate software — GA4 can do it, you just have to set it up:
-
1
Segment your sources
In the session source field, filter for these: chatgpt.com, perplexity.ai, gemini.google.com, copilot.microsoft.com.
-
2
Keep organic separate
google / organic and bing / organic — this is your baseline.
-
3
Tie it to the booking
Have your booking engine send a purchase event to GA4 — without this you only measure visitors, not bookings.
-
4
Review monthly, as a trend
A week's data is noise. The real signal is the monthly trend: is AI traffic and brand search growing?
Methodology — so you know exactly what we measured
The table shows 30 days of Google Analytics 4 data (mid-August to mid-September 2026) from three Hungarian hotels measured in our system. As AI traffic we counted visits with a session source of chatgpt.com, perplexity.ai, gemini.google.com and copilot.microsoft.com; as organic, the google/organic and bing/organic traffic. Important: clicks arriving from Google AI Overview and AI Mode get a google/organic label in the measurement, so the "organic" column actually contains Google-AI-originated visitors too; the two groups can't be separated with today's analytics. Booking = the purchase event sent by the booking engine. We deliberately don't disclose the hotels' names — the pattern is what's interesting, not the players. The sample is small (782 AI sessions), so the rates should be read as ranges, not to the decimal.
What to realistically expect — 3, 6, 12 months
Based on our own fleet's trends, a realistic trajectory looks like this. In the first months, setting up the measurement and sorting out basic citability brings visible change in mentions; in traffic, barely any yet. Around six months, the share of AI traffic creeps upward (we have a hotel where 9% of search traffic is already AI-originated), and the first bookings clearly tied to AI appear. Over a one-year horizon, the biggest item is the rise in brand searches — the channel into which the trust earned in AI flows.
What you shouldn't expect: that AI traffic will replace organic in the short term. It doesn't replace it — it gets ahead of it. An ever larger part of the guest's decision is made before they reach your website.
What to do about it today
By our numbers, AI is today a few per cent of hotel search traffic, with a lower booking rate on the first visit, yet with fast-growing and underestimated influence. So there's no cause for panic nor for a dismissive shrug, but rather for claiming position early: assess where you stand, improve your citability, and follow the trend. When the curve steepens (and every sign says it will), you'll want to already be inside the answers, not just getting started.
Mentio measures this entire chain for hotels: AI mentions, Google AI surfaces, AI traffic and booking conversion, all in one place. The numbers above come from exactly this system.
The biggest shift of the past few years is that search itself has changed: guests no longer necessarily choose from ten blue links — they ask an AI model where to stay. ChatGPT, Claude, Gemini, Perplexity — and now Google itself, with AI Mode and AI Overviews. And the answer is no longer a list of results, but two or three specific hotel names: a decision already made.
Frequently asked questions
Why is the booking rate of AI traffic lower?+
Because the AI-referred guest is typically still researching: comparing, looking around, and often booking later — returning by searching for the hotel's name or arriving directly. In the measurement, that booking is no longer credited to AI — which is why the measured rate is a lower bound.
How large is AI traffic today at a Hungarian hotel?+
In our fleet, roughly 5-9% of search traffic arrives from AI assistants, varying by hotel — and the trend is rising. Small, but no longer negligible, and measurable.
How do I measure AI traffic at my own hotel?+
In GA4, filter by session source: chatgpt.com, perplexity.ai, gemini.google.com, copilot.microsoft.com. Tie it to your booking engine's purchase event, and view it as a monthly trend. The article describes it step by step.
Why can't I see visitors from Claude and Google AI?+
For two reasons. Claude typically doesn't send referral data, so visits from there appear as direct traffic. And guests clicking through from Google AI Overview and AI Mode get a google/organic label, meaning they hide within your organic numbers. That's why source-based AI measurement is always a lower bound.
Read these next
- Google AI Overview: is the age of the Google SERP over? — what every hotel needs to knowOur latest fleet-wide measurement shows that two-thirds of hotel searches now greet guests with an AI summary — and most hotels aren't in…
- Why Can't AI See Your Website? The JavaScript Trap — and How to Avoid ItYour site might be beautiful — but if the content is built with JavaScript, most AI bots see a half-empty page. With a fresh measurement…
- AI visibility for hotels: how to get recommended by ChatGPT, Gemini and PerplexityMore and more guests ask AI assistants for accommodation recommendations. What do AI visibility and GEO mean for hotels — and what concrete…